RBI Governor: AI Can Approve Loans Humans Would Reject

What happened

The head of the Reserve Bank of India (RBI) has said artificial intelligence can approve loans that human loan officers would otherwise reject, framing AI as a capability for lenders to actively harness rather than a threat to be merely contained or restricted. The remarks, reported by TechRadar, signal that India's banking regulator is encouraging financial institutions to expand their use of AI in credit decisioning rather than treating the technology purely as a compliance risk.

Alongside the endorsement of AI's potential to widen access to credit, the message carries a clear accountability expectation: banks deploying AI in lending decisions cannot simply attribute an outcome to "the model" when a customer is declined or affected. Responsibility for the decision remains with the institution, not the algorithm.

Why it matters

This is a notable regulatory signal in a sector where AI adoption has often been slowed by caution around explainability, fairness and liability. By explicitly framing AI as an opportunity to extend credit to applicants who might be turned down under conventional underwriting, the RBI is nudging banks toward broader AI deployment in a core, high-stakes customer decision — one that directly shapes financial inclusion outcomes.

For leaders in digital transformation and experience design, the pairing of "harness the capability" with "you still own the decision" is the more consequential part of the story. It suggests regulators are moving toward a model where AI-driven decisioning is permitted and encouraged, provided institutions maintain clear lines of accountability, explainability and recourse for the humans affected by those decisions.

The Renascence take

The interesting tension here isn't AI-versus-human accuracy — it's who absorbs responsibility when an algorithm's judgement replaces a loan officer's. That's a service-design and trust problem as much as a technology one.

Most coverage of this story will focus on AI "outperforming" human underwriters, but the more important line is the one banks will be tempted to ignore: you cannot outsource accountability to a model. Customers don't experience an algorithm — they experience a decision, an explanation (or lack of one), and a path to appeal. Any lender expanding AI in credit approval needs to invest as heavily in explainability and human escalation routes as it does in the model itself, or it will simply trade one kind of bias for another, less accountable one.